deploy VM Instance using deployment manager – CentOS

You can take a look at this example explaining how to create two VMs and a custom network with Google Deployment Manager. In this repository, you can also find other examples that might be useful.

Anyway, for your specific problem : to create a custom VPC network, use :

- name: my-custom-network
  type: compute.v1.network
  properties:
    routingConfig:
      routingMode: REGIONAL
    autoCreateSubnetworks: true

Check the REST reference of network resource to find all possible parameters.

To reference your custom network and also to use SSD instead of Standard disk for your VM :

1. reference the custom network in networkInterfaces section (network: $(ref.my-custom-network.selfLink))
2. add the disks.initializeParams.diskType property in your VM configuration

To conclude, for your VM, you should write :

- name: vm
  type: compute.v1.instance
  properties:
    zone: northamerica-northeast1-a
    machineType: zones/northamerica-northeast1-a/machineTypes/f1-micro
    disks:
    - deviceName: boot
      type: PERSISTENT
      boot: true
      autoDelete: true
      initializeParams:
        sourceImage: projects/centos-cloud/global/images/family/centos-7
        diskType: https://www.googleapis.com/compute/v1/projects/{{ env["project"] }}/zones/northamerica-northeast1-a/diskTypes/pd-ssd
    networkInterfaces:
      - network: $(ref.my-custom-network.selfLink)
        accessConfigs:
        - name: External NAT
          type: ONE_TO_ONE_NAT

For the disk, notice the pd-ssd at the end of the disks[0].initializeParams.diskType property, instead of pd-standard (default). Check the REST reference of instance resource for other parameters.